Conventionally, in an imaging device such as a digital camera, an imaging element having a plurality of pixels for converting incident light to an electric signal is installed. These plurality of pixels switch the conversion mode to the electric signal on the basis of the incident light intensity and more in detail, switches the linear conversion mode for linearly converting the incident light to an electric signal and the logarithm conversion mode for logarithmically converting it. Further, at the later stage of the imaging element, a signal processing unit for performing characteristic conversion for converting the electric signal obtained by the logarithm conversion mode to a state obtained by the linear conversion mode or converting the electric signal obtained by the linear conversion mode to a state obtained by the logarithm conversion mode is installed, thus all the electric signals are unified to a state obtained by the linear conversion mode or logarithm conversion mode and the processing of the electric signals is simplified.
According to such an imaging element, compared with an imaging element for performing only the linear conversion mode, the timing range of an electric signal is extended, so that even if an object having a wide brightness range is imaged, all the brightness information can be expressed by an electric signal.
On the other hand, the plurality of pixels aforementioned have variations in the I/O characteristic due to differences between the pixels. Therefore, as a method for canceling such variations, there is a method available for correcting output from each pixel and making it coincide with a reference output value (for example, refer to Patent Documents 1 and 2).
Patent Document 1: Japanese Patent Application Hei 11-298799
Patent Document 2: Japanese Patent Application Hei 5-30350